The Fed's hawkish perspective: AI will raise the neutral interest rate in the medium to long term, but will be difficult to alleviate inflationary pressures in the short term.
2026-05-29 20:10:27
In the current macroeconomic cycle in 2026, the US economy is at a crossroads, fraught with a mix of "technology frenzy" and "inflationary problems."
St. Louis Federal Reserve President Alberto Musalem and Federal Reserve official Brenton Schmid have recently expressed extremely hawkish views.
The two officials' policies both point to the same contradiction: the market has high hopes for high productivity in the long term driven by technologies such as artificial intelligence (AI), but the current macroeconomic reality is that inflation is high, long-term inflation expectations continue to rise, and the actual policy interest rate is running below the long-term neutral interest rate, making the tightening of monetary policy clearly insufficient.
Faced with the dual pressures of supply and demand, the Federal Reserve's core objective remains steadfast in its policy goal of bringing inflation back to 2%.
The central bank must demonstrate its determination to suppress inflation to the market, rather than anchoring its policies on the vague notion of "future technological dividends."
Currently, there is a widespread optimistic consensus in the market and academia: AI brings about productivity improvements, which is equivalent to an expansion of total social supply. In theory, it can directly reduce the unit output cost of enterprises, thus becoming a "magic bullet" to eliminate inflation.
However, from the perspective of a mature macroeconomic framework, this transmission path is very likely to backfire in the short term.
On the supply side (micro-level effects, macro-level lag): Although micro-level research confirms that AI has indeed reduced costs and increased efficiency in specific scenarios such as coding and administration, Schmid points out that while the popularization of AI will inhibit new hiring by companies, it will not lead to large-scale layoffs (the labor market as a whole is supported by recruitment in the service industry such as healthcare, and remains stable).
However, the transformation of new technologies into total factor productivity (TFP) at the national level has a significant lag, and the micro-level benefits may even be diluted due to the redistribution of production factors across industries. (That is, even if the cost of AI industry can decrease, the wage costs of industries that provide supporting infrastructure for AI will increase, thus diluting the benefits on average.)
On the demand side (overly optimistic expectations, adding fuel to the fire): the expectation of long-term supply expansion has first and foremost evolved into a positive demand shock.
The capital market's frenzied speculation in the AI sector has generated a huge wealth effect, stimulating a sustained expansion of consumer spending.
Meanwhile, the large-scale commissioning of data centers across the country and the competition for computing chips and power resources have directly increased current marginal production costs and total social demand, leading to high wage growth and persistently high raw material prices.
The superposition of imported inflation: While demand is overheated, the real economy is also facing constraints in traditional sectors.
Schmid emphasized that although the United States has improved its ability to withstand energy shocks, domestic energy producers have not expanded production investment, and rising oil prices continue to weaken residents' purchasing power.
This situation, where "supply hasn't arrived but demand is already hot," means that simply relying on future productivity growth to resolve the current inflation problem carries excessive risk for policy operations. If the central bank allows current aggregate demand to overheat, the high-inflation environment will become firmly established.
Technological revolutions not only affect inflation through both supply and demand, but also fundamentally restructure the analytical framework of monetary policy—namely, the pricing of the long-term neutral natural interest rate (R*).
The neutral interest rate refers to the equilibrium real interest rate level when the economy is neither inflating nor contracting.
When disruptive underlying technologies, represented by AI, are implemented and their long-term sustainability is subsequently verified, the market consensus is that future household income and capital output efficiency will rise in the long term.
This will directly stimulate residents' willingness to borrow and consume, as well as enterprises' demand for real investment. The result of the resonance of these two forces is a significant increase in the central equilibrium natural interest rate of the whole society.
If the neutral interest rate level shifts upwards as a whole due to technological advancements, then even if the Federal Reserve maintains the existing nominal interest rate unchanged (such as the 10-year Treasury yield), the real policy rate will become "lower" relative to the new neutral interest rate, and you might even need to raise interest rates to reach the neutral interest rate.
Goldman Sachs previously projected $7.6 trillion in technology capital expenditures, which would ultimately drive up the overall neutral interest rate level in the form of corporate bonds.
Therefore, Schmid judged that the current tightening力度 of US monetary policy is clearly insufficient.
In order to prevent inflation expectations from spiraling out of control and long-term market interest rates from being forced to rise due to risk compensation, the central bank must proactively follow the upward shift of the neutral interest rate and maintain or even further tighten policy constraints.
To clarify the current policy logic, the Federal Reserve frequently compares the current cycle with the information technology revolution of the 1990s.
A review reveals that the regulatory environment in 2026 is far more stringent and complex than during the Greenspan era.
In the 1990s, in response to productivity improvements brought about by the internet boom, the Federal Reserve under Greenspan implemented an opportunistic deinflation strategy.
Although the Federal Reserve delayed raising interest rates at the time because it saw the benefits of supply-side reforms, it did not blindly lower interest rates. Its nominal policy rate was firmly fixed above 5% for a long time, and the real interest rate was even higher than 3%, but in the end, it achieved the result of inflation falling back to 2%.
While maintaining sufficient policy resolve, the country also benefited from the convergence of multiple external dividends: a continuous improvement in the balance of federal fiscal revenue and expenditure, the peace dividend brought by the end of the Cold War, the deepening of globalization leading to the free flow of people and capital, and the Asian financial crisis triggering a sharp drop in commodity prices.
These factors together helped push the core CPI smoothly back down from around 3% to around 2%.
In contrast, the aforementioned external favorable factors for the US economy in 2026 have either completely disappeared or been completely reversed.
Sluggish investment by domestic energy companies and continued pressure from oil prices have led to a decline in the benefits of globalization.
Worse still, the current real policy interest rate is still operating below the long-term neutral level, indicating that the tightening efforts are clearly insufficient; the core inflation reading has not only deviated significantly from the 2% policy target, but long-term inflation expectations are also continuing to rise.
Historical experience shows that the mild inflationary environment during the technological boom of the 1990s is an exceptional case that cannot be replicated.
Given the increasingly entrenched tail risks of high inflation, the lessons learned from the 1970s, which resulted in runaway inflation due to misjudgments of productivity and potential growth ceilings, are more cautionary than the experiences of the 1990s.
The Federal Reserve must not repeat the mistakes of the past and must take a tougher stance in response to the current macroeconomic challenges.
Based on all the above deductions, both Musalaim's stringent requirements for a closed-loop macroeconomic data system and Schmid's characterization of inflation as a non-temporary surge point to a clear asset pricing conclusion: the US macroeconomy is entering a long-term plateau period characterized by "high interest rates and high constraints."
The threshold for interest rate cuts has been significantly raised: In the short term, any micro-level AI benefits or localized cost reductions and efficiency improvements cannot be directly converted into a catalyst for interest rate cuts.
The Fed's policy shift (the inflection point of interest rate cuts) has only one prerequisite: seeing solid macroeconomic data (such as a slowdown in unit labor costs and a continuous decline in core PCE) confirming that supply expansion has outpaced speculative demand, driving inflation to steadily converge to 2%.
The window for further tightening is not closed: Given that the current policy tightening is not sufficient and long-term inflation expectations continue to rise slowly, the Federal Reserve needs to consider how to further implement policy constraints.
This not only means that nominal policy rates will remain high for a longer period of time, but also, as Schmid suggested, the Federal Reserve may continue to withdraw liquidity beyond the short end by actively adjusting its balance sheet (quantitative tightening), thus strengthening the tightening effect of the policy.
The central level of long-term interest rates is trending upward: the technological boom is pushing up the equilibrium natural interest rate, coupled with long-term structural pressures from fiscal and energy costs, which means that even if inflation falls in the future and the Federal Reserve begins to normalize its policies, the room for interest rate cuts will be extremely limited.
The "ultra-low interest rate era" that we have become accustomed to over the past decade is unlikely to return, and the central level of long-term US Treasury yields will remain at a significantly high level in the coming years.
From the current perspective, maintaining a highly prudent policy stance and taking the restoration of price stability as the core focus are the highest consensus reached among the hawks within the Federal Reserve.
Only by thoroughly suppressing inflation through tightening policies can a healthy macroeconomic foundation be provided for the long-term development of fundamental technologies such as artificial intelligence. Before that, any overly optimistic market expectations regarding looser policies will face the harsh reality of a changing macroeconomic environment.
St. Louis Federal Reserve President Alberto Musalem and Federal Reserve official Brenton Schmid have recently expressed extremely hawkish views.
The two officials' policies both point to the same contradiction: the market has high hopes for high productivity in the long term driven by technologies such as artificial intelligence (AI), but the current macroeconomic reality is that inflation is high, long-term inflation expectations continue to rise, and the actual policy interest rate is running below the long-term neutral interest rate, making the tightening of monetary policy clearly insufficient.
Faced with the dual pressures of supply and demand, the Federal Reserve's core objective remains steadfast in its policy goal of bringing inflation back to 2%.
The central bank must demonstrate its determination to suppress inflation to the market, rather than anchoring its policies on the vague notion of "future technological dividends."
Ideal vs. Reality: The Complex Relationship Between AI Development and Inflation
Currently, there is a widespread optimistic consensus in the market and academia: AI brings about productivity improvements, which is equivalent to an expansion of total social supply. In theory, it can directly reduce the unit output cost of enterprises, thus becoming a "magic bullet" to eliminate inflation.
However, from the perspective of a mature macroeconomic framework, this transmission path is very likely to backfire in the short term.
On the supply side (micro-level effects, macro-level lag): Although micro-level research confirms that AI has indeed reduced costs and increased efficiency in specific scenarios such as coding and administration, Schmid points out that while the popularization of AI will inhibit new hiring by companies, it will not lead to large-scale layoffs (the labor market as a whole is supported by recruitment in the service industry such as healthcare, and remains stable).
However, the transformation of new technologies into total factor productivity (TFP) at the national level has a significant lag, and the micro-level benefits may even be diluted due to the redistribution of production factors across industries. (That is, even if the cost of AI industry can decrease, the wage costs of industries that provide supporting infrastructure for AI will increase, thus diluting the benefits on average.)
On the demand side (overly optimistic expectations, adding fuel to the fire): the expectation of long-term supply expansion has first and foremost evolved into a positive demand shock.
The capital market's frenzied speculation in the AI sector has generated a huge wealth effect, stimulating a sustained expansion of consumer spending.
Meanwhile, the large-scale commissioning of data centers across the country and the competition for computing chips and power resources have directly increased current marginal production costs and total social demand, leading to high wage growth and persistently high raw material prices.
The superposition of imported inflation: While demand is overheated, the real economy is also facing constraints in traditional sectors.
Schmid emphasized that although the United States has improved its ability to withstand energy shocks, domestic energy producers have not expanded production investment, and rising oil prices continue to weaken residents' purchasing power.
This situation, where "supply hasn't arrived but demand is already hot," means that simply relying on future productivity growth to resolve the current inflation problem carries excessive risk for policy operations. If the central bank allows current aggregate demand to overheat, the high-inflation environment will become firmly established.
Displacement of the long-term anchor: The relationship between technological development and the neutral interest rate
Technological revolutions not only affect inflation through both supply and demand, but also fundamentally restructure the analytical framework of monetary policy—namely, the pricing of the long-term neutral natural interest rate (R*).
The neutral interest rate refers to the equilibrium real interest rate level when the economy is neither inflating nor contracting.
When disruptive underlying technologies, represented by AI, are implemented and their long-term sustainability is subsequently verified, the market consensus is that future household income and capital output efficiency will rise in the long term.
This will directly stimulate residents' willingness to borrow and consume, as well as enterprises' demand for real investment. The result of the resonance of these two forces is a significant increase in the central equilibrium natural interest rate of the whole society.
If the neutral interest rate level shifts upwards as a whole due to technological advancements, then even if the Federal Reserve maintains the existing nominal interest rate unchanged (such as the 10-year Treasury yield), the real policy rate will become "lower" relative to the new neutral interest rate, and you might even need to raise interest rates to reach the neutral interest rate.
Goldman Sachs previously projected $7.6 trillion in technology capital expenditures, which would ultimately drive up the overall neutral interest rate level in the form of corporate bonds.
Therefore, Schmid judged that the current tightening力度 of US monetary policy is clearly insufficient.
In order to prevent inflation expectations from spiraling out of control and long-term market interest rates from being forced to rise due to risk compensation, the central bank must proactively follow the upward shift of the neutral interest rate and maintain or even further tighten policy constraints.
A Mirror from History: A Comparison of the Regulatory Environments During Two Periods of Technological Development
To clarify the current policy logic, the Federal Reserve frequently compares the current cycle with the information technology revolution of the 1990s.
A review reveals that the regulatory environment in 2026 is far more stringent and complex than during the Greenspan era.
In the 1990s, in response to productivity improvements brought about by the internet boom, the Federal Reserve under Greenspan implemented an opportunistic deinflation strategy.
Although the Federal Reserve delayed raising interest rates at the time because it saw the benefits of supply-side reforms, it did not blindly lower interest rates. Its nominal policy rate was firmly fixed above 5% for a long time, and the real interest rate was even higher than 3%, but in the end, it achieved the result of inflation falling back to 2%.
While maintaining sufficient policy resolve, the country also benefited from the convergence of multiple external dividends: a continuous improvement in the balance of federal fiscal revenue and expenditure, the peace dividend brought by the end of the Cold War, the deepening of globalization leading to the free flow of people and capital, and the Asian financial crisis triggering a sharp drop in commodity prices.
These factors together helped push the core CPI smoothly back down from around 3% to around 2%.
In contrast, the aforementioned external favorable factors for the US economy in 2026 have either completely disappeared or been completely reversed.
Sluggish investment by domestic energy companies and continued pressure from oil prices have led to a decline in the benefits of globalization.
Worse still, the current real policy interest rate is still operating below the long-term neutral level, indicating that the tightening efforts are clearly insufficient; the core inflation reading has not only deviated significantly from the 2% policy target, but long-term inflation expectations are also continuing to rise.
Historical experience shows that the mild inflationary environment during the technological boom of the 1990s is an exceptional case that cannot be replicated.
Given the increasingly entrenched tail risks of high inflation, the lessons learned from the 1970s, which resulted in runaway inflation due to misjudgments of productivity and potential growth ceilings, are more cautionary than the experiences of the 1990s.
The Federal Reserve must not repeat the mistakes of the past and must take a tougher stance in response to the current macroeconomic challenges.
Final Assessment: Outlook on Future Interest Rate Trends
Based on all the above deductions, both Musalaim's stringent requirements for a closed-loop macroeconomic data system and Schmid's characterization of inflation as a non-temporary surge point to a clear asset pricing conclusion: the US macroeconomy is entering a long-term plateau period characterized by "high interest rates and high constraints."
The threshold for interest rate cuts has been significantly raised: In the short term, any micro-level AI benefits or localized cost reductions and efficiency improvements cannot be directly converted into a catalyst for interest rate cuts.
The Fed's policy shift (the inflection point of interest rate cuts) has only one prerequisite: seeing solid macroeconomic data (such as a slowdown in unit labor costs and a continuous decline in core PCE) confirming that supply expansion has outpaced speculative demand, driving inflation to steadily converge to 2%.
The window for further tightening is not closed: Given that the current policy tightening is not sufficient and long-term inflation expectations continue to rise slowly, the Federal Reserve needs to consider how to further implement policy constraints.
This not only means that nominal policy rates will remain high for a longer period of time, but also, as Schmid suggested, the Federal Reserve may continue to withdraw liquidity beyond the short end by actively adjusting its balance sheet (quantitative tightening), thus strengthening the tightening effect of the policy.
The central level of long-term interest rates is trending upward: the technological boom is pushing up the equilibrium natural interest rate, coupled with long-term structural pressures from fiscal and energy costs, which means that even if inflation falls in the future and the Federal Reserve begins to normalize its policies, the room for interest rate cuts will be extremely limited.
The "ultra-low interest rate era" that we have become accustomed to over the past decade is unlikely to return, and the central level of long-term US Treasury yields will remain at a significantly high level in the coming years.
From the current perspective, maintaining a highly prudent policy stance and taking the restoration of price stability as the core focus are the highest consensus reached among the hawks within the Federal Reserve.
Only by thoroughly suppressing inflation through tightening policies can a healthy macroeconomic foundation be provided for the long-term development of fundamental technologies such as artificial intelligence. Before that, any overly optimistic market expectations regarding looser policies will face the harsh reality of a changing macroeconomic environment.
- Risk Warning and Disclaimer
- The market involves risk, and trading may not be suitable for all investors. This article is for reference only and does not constitute personal investment advice, nor does it take into account certain users’ specific investment objectives, financial situation, or other needs. Any investment decisions made based on this information are at your own risk.
